north african hybrid sparrows (passer domesticus p...

North African hybrid sparrows (Passer domesticus, P.hispaniolensis) back from oblivion – ecological segregationand asymmetric mitochondrial introgression betweenparental speciesAbdelkrim Ait Belkacem1, Oliver Gast2,3, Heiko Stuckas2, David Canal4, Mario LoValvo5,Gabriele Giacalone6 & Martin P€ackert2

1Faculty of Sciences of nature and lifes Department of Agropastoralism, University of Djelfa, BP. 3117, 17000 Djelfa, Algeria2Senckenberg Naturhistorische Sammlungen, K€onigsbr€ucker Landstraße 159, D-01109 Dresden, Germany3Institute of Vertebrate Biology, Czech Academy of Sciences, External Research Facility Studenec, Studenec 122, 675 02, Kon�e�s�ın, Czech Republic4Department of Evolutionary Ecology, Estaci�on Biol�ogica de Do~nana – CSIC, Avda. Am�erico Vespucio s/n, 41092 Seville, Spain5Dipartimento di Scienze e Tecnologie Biologiche, Chimiche e Farmaceutiche, Via Archirafi 18, I-90123 Palermo, Italy6Cooperativa Silene, Via Dondes Regio, 8/a, 90127 Palermo, Italy

Keywords

Agricultural landscape mosaic, Algeria,

breeding phenology, NADH dehydrogenase,

nest site choice.

Correspondence

Martin P€ackert, Senckenberg Naturhistorisch

Sammlungen, Landstraße 159,

K€onigsbr€ucker, D-01109 Dresden, Germany.

Tel: 49(0)351-7958414344;

Fax: 49(0)351-7958414327;

E-mail: [email protected]

Funding Information

No funding information provided.

Received: 3 February 2016; Revised: 31 May

2016; Accepted: 3 June 2016

Ecology and Evolution 2016; 6(15): 5190–

5206

doi: 10.1002/ece3.2274

Abstract

A stabilized hybrid form of the house sparrow (Passer domesticus) and the

Spanish sparrow (P. hispaniolensis) is known as Passer italiae from the Italian

Peninsula and a few Mediterranean islands. The growing attention for the Ital-

ian hybrid sparrow and increasing knowledge on its biology and genetic consti-

tution greatly contrast the complete lack of knowledge of the long-known

phenotypical hybrid sparrow populations from North Africa. Our study pro-

vides new data on the breeding biology and variation of mitochondrial DNA in

three Algerian populations of house sparrows, Spanish sparrows, and phenotyp-

ical hybrids. In two field seasons, the two species occupied different breeding

habitats: Spanish sparrows were only found in rural areas outside the cities and

bred in open-cup nests built in large jujube bushes. In contrast, house sparrows

bred only in the town centers and occupied nesting holes in walls of buildings.

Phenotypical hybrids were always associated with house sparrow populations.

House sparrows and phenotypical hybrids started breeding mid of March, and

most pairs had three successive clutches, whereas Spanish sparrows started

breeding almost one month later and had only two successive clutches. Mito-

chondrial introgression is strongly asymmetric because about 75% of the rural

Spanish sparrow population carried house sparrow haplotypes. In contrast,

populations of the Italian hybrid form, P. italiae, were genetically least diverse

among all study populations and showed a near-fixation of house sparrow hap-

lotypes that elsewhere were extremely rare or that were even unique for the Ital-

ian Peninsula. Such differences between mitochondrial gene pools of Italian and

North African hybrid sparrow populations provide first evidence that different

demographic histories have shaped the extant genetic diversity observed on

both continents.

Introduction

Hybridization has been considered as the driving force of

speciation in several groups of organisms (review in:

Abbot et al. 2013). In birds, however, hybridization pro-

cesses were most intensely studied in the context of a

breakdown of reproductive barriers in secondary contact

rather than with respect to the emergence of truly stabi-

lized hybrid species. A great number of narrow-range

avian hybrid zones have been circumscribed in the

Palearctic, but in most areas, distribution range of hybrids

is limited to rather narrow zones of secondary contact in

coexistence with both parental species (Haffer 1989; Ali-

abadian et al. 2005). In contrast, genetic evidence of

5190 ª 2016 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use,

distribution and reproduction in any medium, provided the original work is properly cited.

hybrid origin of a bird species is rare, being reliably docu-

mented for a handful of taxa only, for example, for the

Haiwaiian duck (Anas wyvilliana; Lavretsky et al. 2015),

the imperial pheasant (Lophura imperialis; Hennache et al.

2003), Audubon’s warbler (Dendroica aududoni; Brelsford

et al. 2011), and for the Italian sparrow (Passer italiae), a

stabilized hybrid form from the Italian Peninsula. The

Italian sparrow occupies a wide distribution range in

absence of either of its parental species: the house spar-

row (P. domesticus) and the Spanish sparrow (P. hispan-

iolensis; Elgvin et al. 2011; Hermansen et al. 2011, 2014;

Trier et al. 2014). This peculiar sparrow has obtained the

straightforward recognition from the scientific community

worldwide as a homoploid hybrid species (Gill 2014;

Arnold 2016; : p. 17–18), and its proposed species rank

was accordingly accepted by many taxonomic authorities

(e.g., Clements et al. 2015; Dickinson and Christidis

2014). The limited geographic distribution range of the

Italian sparrow is a sideline aspect rendering further cred-

ibility to the suggested species status of the Italian spar-

row. According to recent genetic studies, the hybrid

Italian sparrow would be distributed only throughout the

entire Italian peninsula and in a few insular populations

on Sicily, Sardinia (see maps in Elgvin et al. 2011; Her-

mansen et al. 2011; Trier et al. 2014), Malta, and Crete

(Clement 1999). Despite the existence of wide areas of

sympatry among the two parental species in Europe and

Central Asia, hybrid phenotypes have been rarely docu-

mented in Eurasia north of the Alps. Previous work on

the Iberian Peninsula reported rare hybrid individuals

from areas of local contact (Alonso 1985), but there is

recent evidence for genetic incompatibilities between the

two parental species from a mixed Spanish population

(Hermansen et al. 2014).

The situation on the Eurasian continent is greatly con-

trasted by the distribution pattern of the two – respec-

tively, three – sparrow species south of the

Mediterranean, on the African continent. In fact, a great

number of ornithological papers have documented spar-

row populations that are phenotypically intermediate

between the house sparrow and the Spanish sparrow

throughout a vast region in North Africa (review in

T€opfer 2006: p. 120–121). A first cartographic documen-

tation of putative hybrid sparrow populations in North

Africa by Meise (1936) was later complemented and

improved by contributions from other authors (Johnston

1969; Summer-Smith and Vernon 1972; Metzmacher

1986; Haffer and Hudde 1997). Recent avifaunistic sur-

veys demonstrated that North African “hybrid sparrows”

are “locally highly abundant” with local records from

Algeria accounting “more than one-third of the total

number of individuals of all species inventoried” (Gue-

zoul et al. 2010; Guezoul et al. 2011; Bendjoudi et al.

2013). Thus, the spatial distribution of house and Spanish

sparrows in North Africa resembles a mosaic of allopatric

and sympatric populations with more than 50 occurrence

records of the intermediate phenotype (Fig. 1). Summer-

Smith and Vernon (1972) described the situation in

North Africa as a “spatially diffuse hybrid zone extending

from eastern Algeria through Tunisia into eastern

Figure 1. Distribution of the house sparrow (P. domesticus, dark gray), Spanish sparrow (P. hispaniolensis, dark red-brown), and hybrid

phenotypes (light beige) in North Africa; different symbol sizes were chosen to show local co-occurrence of the three forms; main sources for

occurrence data: Meise (1936), Johnston (1969), Summers-Smith and Vernon (1972), and Metzmacher (1986); our study sites in Algeria: (1)

Hassi-El Euch; (2) Djelfa.

ª 2016 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. 5191

A. Ait Belkacem et al. North African Hybrid Sparrows Ecology and Genetics

Tripolitania” (illustration in Summers-Smith 1988, fig.

44). However, to date, reliable data on life history traits

and the genetic constitution of these North African spar-

row populations are still missing.

Here, we provide field data on the breeding biology

(habitat preferences, phenology, and reproductive success)

of P. domesticus, P. hispaniolensis, and their putative

hybrids from two breeding seasons on two study sites in

Algeria. Throughout their sympatric European range,

breeding phenology and nest site preferences of house

sparrows and Spanish sparrows are partly diverged

(Alonso 1986; Murgui 2011; Tu 2012) and our field study

aims to determine whether such ecological segregation

occurs in North African populations. Field data are

backed by morphometric comparisons among the three

locally coexisting phenotypes and by genetic analysis of

mitochondrial DNA. As is known, the house sparrow

mitogenome has almost completely introgressed into the

Italian hybrid form, P. italiae (about 98% of study indi-

viduals of Hermansen et al. 2011) and we expect to con-

firm such a fixation of P. domesticus haplotypes for North

African hybrids, too. Considering the effective reproduc-

tive barriers between House sparrows, Spanish sparrows

on the Iberian Peninsula (Hermansen et al. 2014), and

between the Italian hybrid form and the parental species

throughout the Italian Peninsula (Trier et al. 2014), we

expect a rather limited degree of mitochondrial introgres-

sion between P. domesticus and P. hispaniolensis in Alge-

ria.

Material and Methods

Study area

Fieldwork and nest surveys were carried out in the years

2011 and 2012 at two locations: Hassi El-Euch (35°90

North; 3°140 East, 910 m. a. s. l) and Djelfa (34°400

North; 3°150 East, 1138 m.a.s.l), in the region of Zehrez,

located at about 275 km in the southeast of Algiers

(Fig. 1). The region is semi-arid with relatively cool win-

ters and precipitations ranging from 217 to 337 mm

(Belkacem et al. 2012). The Institut de Technologie

Moyen Agricole Sp�ecialis�ee (ITMAS) was our study site at

Djelfa. The institute, located in an urban environment,

has an area of ca. 16 ha and consists of buildings and

houses surrounded by agricultural study lots for wheat

production and cattle farming, some of them bordered by

small stands of Aleppo pine (Pinus halepensis). In Hassi-

El Euch, we studied two sparrow populations: the urban

study site was located at the local school and the sur-

rounding quarters that were bordered by fallow land in

the North, whereas the rural study site consisted of large

agricultural fields of wheat covering about 1450-ha extent.

The abundance of jujube bushes – the preferred breeding

habitat of Spanish sparrows – in the area was rather low

(2% of the total plot; see Belkacem et al. 2012).

Fieldwork started with the onset of the breeding period

of house sparrows in mid-March (15th March in both

study periods). Nests were regularly checked to ascertain

laying date, clutch size, hatching date, and number of

fledglings of first, second, and third clutches.

We used general and generalized linear models (GLM)

to explore for differences among populations in laying

dates (Gaussian distribution), clutch sizes (Poisson distri-

bution), and reproductive success (young fledged in rela-

tion to clutch size; binomial distribution). If differences

were significant, post hoc Tukey’s test was applied to test

for paired differences between the three study popula-

tions. In these models, laying date was included as a pre-

dictor because may be an important determinant of

reproductive success in birds (Newton 2008). Further,

year was included as a random effect to account for

between-year variability in environmental conditions

influencing reproductive success. Selection of the mini-

mum adequate models was conducted by sequentially

dropping nonsignificant terms from fully saturated mod-

els (containing all main effects and interactions) in a hier-

archical way, starting with the least significant order

terms. We systematically performed model diagnostics

statistics while modeling to avoid misleading conclusions

based on statistical artifacts. As we performed multiple

tests using the phenologic data set, we applied the Ben-

jamini–Yekutieli correction for multiple tests (which is

more appropriate than other methods of correction for

multiple comparison tests; Narum 2006).

Morphology

Phenotypical assignment of individual birds to either of

the two sparrow species was based on diagnostic plumage

traits (Fig. S1). These refer to the color pattern of the

crown, the nape, the cheek and the back, the extent of

black chest patch, and the presence and absence of black-

streaked pattern on the flanks (following Summers-Smith

1988; Cramp and Perrins 1994). Phenotypic hybrid indi-

viduals were identified according to their intermediate

plumage characters, particularly with respect to crown

color, cheek color, and black spots on the flanks and

underparts (Fig. S1). To take into account the consider-

able variation among hybrid phenotypes, different meth-

ods of quantifying a hybrid index have been established

(Meise 1936; Johnston 1969; Lo Valvo and Lo Verde

1987) and applied for the classification of hybrid pheno-

types across an alpine contact zone between P. domesticus

and the hybrid form P. italiae (Hermansen et al. 2011).

Because all these hybrid indices were based on estimated


North African Hybrid Sparrows Ecology and Genetics A. Ait Belkacem et al.

numerical values that had never been quantified or mea-

sured (see T€opfer 2006), they are all highly subjective and

not directly comparable. Therefore, for our analyses of

Algerian populations, we classified all intermediate pheno-

types as phenotypical hybrids opposed to the 100% par-

ental domesticus and hispaniolensis phenotypes (compare

Fig. S1). In the context of the present study, the use of

this binomial method may have led to a slight underesti-

mation of phenotypic hybrid numbers. Thus, our results

likely represent a conservative estimate of the actual situa-

tion in Algerian populations.

Although distinction of the two sparrow species is

easy in males, females of both species are almost indis-

tinguishable from a distance (Cramp and Perrins 1994).

Generally, females of both species are very much alike,

but in comparison, Spanish sparrows have somewhat

paler crown and face, distinct dusky furrows at the

flanks, a whiter belly and sharper streaks on mantle

(Cramp and Perrins 1994; Haffer and Hudde 1997).

However, none of these characters is really diagnostic

and, therefore, in cases of doubt, we finally classified

females according to the phenotype of the males they

were paired with (in accordance with Hermansen et al.

(2011), who classified females according to “geographical

location and the corresponding male phenotype”). This

classification is naturally prone to error because it a pri-

ori neglects any kind of female mating preference, how-

ever, for the phenotypical classification of females there

is barely an alternative.

We used of 245 individuals for morphometric analysis.

Birds were weighed (to the nearest 0.1 g) and measured

for body length (to the nearest 0.5 mm), from the tip of

the bill to the tip of the longest tail feather. Wing span

was measured as the distance between wing tips with both

wings stretched, tarsus length as the distance between the

tarsal joint and the metatarsal joint, and beak length was

measured from the skull to the tip of the bill (bill-to-

skull; Eck et al. 2011). To analyze differences between

species and sexes, we ran GLM for each trait of interest

(i.e., body mass, wing span, tarsus, and beak length)

including the interaction between species and sex as a

predictor. If differences were significant, post hoc Tukey’s

test was applied to test for paired differences in the traits

between species (P. domesticus, P. hispaniolensis, and

hybrids). In addition, we also run a model using the

scores from a principal component analysis (PCA; using

both the covariance and the correlation matrix approach)

based on the morphometric measures detailed above.

PCA was calculated including and excluding mass, as this

trait is not structural and may largely vary along the day

and breeding stage. Results using both sets of variables

were qualitatively similar, and therefore, we only show

here data from PCAs excluding mass to avoid unnecessary

repetition. Selection of the minimum adequate models

was as described above.

Statistical analyses on the breeding biology and mor-

phometry of sparrows were implemented in R 3.1.2 (R

Development Core Team 2015) and SPSS 14.0.

Genetic analysis

We used a total of 185 samples belonging to eleven popu-

lations of house sparrows and Spanish sparrows. Samples

are stored at the facilities of the Senckenberg Natural His-

tory Collection Dresden in 95% ethanol or thymol tissue-

buffer at �85°C until further use.

In Algeria, we collected samples from three populations

in two localities: two populations constituted by house

sparrows and phenotypical hybrids (Djelfa: n = 43, 28

P. domesticus + 15 hybrids; urban, Hassi El-Euch: n = 19,

10 P. domesticus + 9 hybrids) and one constituted by

Spanish sparrows (rural, Hassi El-Euch: n = 26). Due to

the crucial assignment of female phenotypes, we included

only male specimens into the analysis. Algerian popula-

tions were compared (1) with populations from continen-

tal Europe and North Africa of either of the two parental

species P. hispaniolensis (Sevilla, Spain, n = 22; Egypt,

n = 7) and P. domesticus (Dresden, Germany, n = 17;

Morocco, n = 6) and (2) with populations of P. italiae

from Sicily and Mediterranean islands (Fraginesi, W

Sicily, n = 10; Maletto, E Sicily, n = 11; Ustica island,

n = 9; Lampedusa island, n = 15). Although being pheno-

typically rather similar to P. hispaniolensis, these popula-

tions from Sicily and neighboring islands are usually

thought to represent the Italian stabilized hybrid form.

The population from Lampedusa is mostly neglected the

recent taxonomic standard literature, and when discussed,

it is rather affiliated to the Spanish sparrow, P. hispan-

iolensis (Massetti 2009). However, like on other Mediter-

ranean islands and similar to sparrows from Sicily, the

birds in our study population from Lampedusa were phe-

notypically close to P. hispaniolensis but completely lacked

the typical black breast and flank stripes (compare

Fig. S2). Therefore, we also classified the Lampedusa pop-

ulation as belonging to the hybrid form P. italiae.

To estimate the dimension of introgression of parental

mitochondrial lineages into North African hybrid popula-

tions, we chose the NADH dehydrogenase subunit 2

(ND2) as a marker gene for comparison with sequence

data available from previous studies (Elgvin et al. 2011;

Hermansen et al. 2011). Mitochondrial ND2 lineages of

house sparrows and Spanish sparrows are strongly

diverged, but local gene pools of the Italian hybrid form

P. italiae are composed to near 100% of house sparrow

haplotypes (Hermansen et al. 2011) and we would expect

a similar situation in North African hybrids.



For comparison with our own sequence data set, we

added 47 ND2 sequences from two further European

populations studied in Hermansen et al. (2011): Norwe-

gian P. domesticus from Olso (GenBank numbers:

JN090466–JN090481). P. hispaniolensis from Sardinia

(GenBank numbers: JN090483–JN090497), and P. italiae

from Central Italy (GenBank numbers: JN90498–JN90512).

DNA was extracted from tissue samples in a standard

chloroform–isoamylalcohol extraction using dode-

cyltrimethylammonium bromide (DTAB) that followed a

modified protocol for blood samples (Gustincich et al.

1991). For amplification of the entire 1079-bp long ND2

gene, we used a pair of external primer pairs: H6313 (50-ACT CTT RTT TAA GGC TTT GAA GGC-30) and L5216

(50-GGC CCA TAC CCC GRA AAT G-30). The poly-

merase chain reaction (PCR) set-up consisted of 25 lLtotal volume each containing 2.5 lL 109 PCR buffer,

0.5 lL dNTPs (10 mM), 1.0 lL of each primer

(10 pmol), 0.2 lL Taq polymerase, 20–50 ng template

DNA, and ddH2O to yield a final volume of 25 lL.Negative controls containing no DNA template were run

with each PCR to rule out an effect of possible contami-

nation. We empirically determined the optimal annealing

temperature of 57°C for our PCR primer pair in a gradi-

ent PCR. Except for that little modification, the PCR pro-

tocol followed that of Hermansen et al. (2011).

Because we received suboptimal sequencing results

when using one of the external primers H6313 or L5216

for a considerable amount of samples, we designed two

internal sequencing primers using OLIGOANALYZER

1.03 Software (Owczarzy et al. 2008) and a preliminary

ND2 alignment including sequences from both species:

forward primer PasserND2_seqintF, 50-ACC ATC ACT

AAA TCC CAC ACT C-30; reverse primer PasserND2_se-

qintR 50-TAA GGT GAG GAA GAC TGT TGA G-30). We

used the BigDye� Terminator Cycle Sequencing Kit

(Applied Biosystems Inc., Darmstadt, Germany) to

sequence the ND2 PCR products. Sequencing reactions

(Vf = 10 lL) consisted of 50–100 ng DNA template per

1000 bp (in most cases 1 lL PCR product), 1 lL Primer

(5 pmol), 0.5 lL BigDye, 2.25 lL 59 sequencing buffer,

and 5.25 lL ddH2O. Sequencing PCR products were

purified using SephadexTM purification. Sequencing analy-

sis of PCR products was performed on an ABI3730xl cap-

illary sequencer (Applied Biosystems).

All 145 ND2 sequences (own samples and GenBank

sequences, 717 bp length) were manually aligned using

MEGA 5.1 (Tamura et al. 2011). Variable and ambiguous

sites were visually checked for accuracy and validated by

examining the raw data electropherogram output file.

Nucleotide sequences were translated into protein

sequences with MEGA 5.1 in order to control for stop

codons and thus to exclude numt (nuclear mitochondrial

DNA) sequences as a potential source of error. For

genetic analyses, the alignments were cut at the end of

both strands and sequences with missing data or ambigu-

ous sites were eliminated. ND2 sequence data set was

deposited at GenBank under accession numbers

KX370619-KX370815.

We reconstructed a minimum-spanning haplotype net-

work using TCS 1.2.1, phylogenetic network estimation

using statistical parsimony (Clement et al. 2000), and cal-

culated genetic diversity indices with DnaSP 5.10 (Librado

and Rozas 2009). DnaSP was also used to infer the mis-

match distributions under a model of population growth

and decline (Rogers and Harpending 1992; Rogers 1995).

Expected distributions were calculated by a priori estimat-

ing theta initial and tau from the original sequence data

for each population separately. To test for deviations

from neutrality, we calculated Tajima’s D (Tajima, 1989)

and performed the McDonald–Kreitman test (McDonald

and Kreitman 1991) with DnaSP V. 5.10 (Librado and

Rozas 2009). To compare synonymous and nonsynony-

mous variation within and between species with the

McDonald–Kreitman test, we used reduced sequence data

sets of allopatric house sparrows and Spanish sparrows.

Italian sparrows could generally not be tested against

house sparrows due to their very similar mitochondrial

gene pools (the contingency table could not be com-

puted) and were excluded from this analysis.

Results

Habitat choice and breeding phenology

We surveyed a total of 103 nests in the Djelfa population

(58 nests in 2011 and 45 nests in 2012), whereas, in

Hassi-El Euch, we surveyed 74 nests in the urban area (43

nests in 2011 and 31 nests in 2012) and 251 nests in the

rural population (124 nests in 2011 and 127 nests in

2012). The two urban Algerian populations of Djelfa and

Hassi El-Euch were constituted by house sparrows, and

phenotypical hybrid sparrows were exclusively associated

with that species (“mixed urban populations” hereafter).

In contrast, Spanish sparrows occurred only in the rural

agricultural areas in the surroundings of the human set-

tlements. In the rural surroundings of Djelfa, where

jujube bushes were completely absent, no breeding Span-

ish sparrows were recorded during the two field seasons.

In contrast, large breeding colonies existed throughout

the Hassi-El Euch area despite the low abundance of

jujube bushes (<2% of the rural plots). In the urban

areas, the great majority of house sparrow and phenotypi-

cal hybrid nests were located in holes of building walls

(Fig. 2A and C; at Djelfa 90% of all nests in 2011 and



http://www.ncbi.nlm.nih.gov/nuccore/JN090466






87% in 2012). All other nests were found in the vicinity

or at human buildings, such as power poles. In contrast,

all nests of the Spanish sparrows were built in colonies in

large jujube bushes (Zyziphus lotus) along large wheat

(Triticum durum) fields (Fig. 2B and D).

Breeding dates were similar between years (F = 1.575,

P = 0.21), but consistently differed between populations

(F = 5.523, P = 0.004) in the two study years (F = 0.329,

P = 0.719). Post hoc analyses showed that the breeding

dates were similar between the two mixed urban popula-

tions (t = �1.532, P = 0.273) and thus were simplified as

a unique “population.” After that, Tukey’s test confirmed

that the P. hispaniolensis population breeds later than

P. domesticus (t = 2.94, P = 0.003). Analyzed by clutches,

P. hispaniolensis started the first clutch later than the two

mixed urban populations by mid of April (at Djelfa:

t = 43.99, P < 0.001; Hassi El-Euch: t = 41.21, P < 0.001)

and both mixed populations had similar breeding dates

with first clutches laid during second half of March

(t = 0.49, P = 0.87). In the second clutch, the mixed

urban population at Djelfa bred significantly earlier than

the mixed population at Hassi El-Euch (t = �4.97,

P < 0.001) and both of them, significantly earlier than

the P. hispanonlensis population (Djelfa: t = �44.742,

P < 0.001; Hassi El-Euch: t = �42.516, P < 0.001).

Breeding dates in the third clutch were similar for the

two mixed urban populations (t = �1.404, P = 0.166).

P. hispaniolensis did not lay a third clutch, and thus, their

breeding period was shorter than that of neighboring

house and hybrid sparrow populations. Breeding phenol-

ogy of the two species and their hybrids was not entirely

synchronized, because in both years laying periods of

Spanish sparrows often coincided with incubation and

parental care of house sparrows and hybrids (Fig. 3).

(A) (B)

(C) (D)Figure 2. Nesting sites of house sparrows,

Spanish sparrows in Algeria, and phenotypical

hybrids; (A, C) P. domesticus: nest with eggs

and burrows in brick wall, both at Djelfa –

phenotypical hybrids show the same nesting

site preference; (B, D) P. hispaniolensis: nest

with eggs and breeding colony in jujube

bushes, both at Hassi El-Euch (photos: A. Ait

Belkacem).

Figure 3. Breeding phenology of Algerian

house sparrows (P. domesticus), Spanish

sparrows (P. hispaniolensis), and their hybrids

in the two study seasons; laying dates are

numbered such that day 78 is March 19th and

so on; y-axis: percentage of total clutches of a

species per season; laying periods of 1st and

2nd clutches of Spanish sparrows marked with

gray shade; plateaus (with no increase of

clutches initiated) indicate periods of

incubation and parental care; note the

synchronization of house sparrows’ and

Spanish sparrows’ incubation and laying

periods toward the end of the breeding season

2011 (less synchronized in the following year).



Laying periods of the two species coincided most closely

at the end of the breeding season, thus during second

clutches of Spanish sparrows and third clutches of house

sparrows and hybrids (Fig. 3).

Overall, clutch sizes and number of hatched chicks,

which were unaffected by the breeding date (GLMM:

Z = �0.056, P = 0.57 and Z = �0.504, P = 0.614, for egg

and hatched chicks, respectively), were similar in the

Spanish sparrow population and the two mixed urban

populations (Table 1), which was confirmed in the analy-

ses by clutch (all comparisons for first, second, and third

clutches: P > 0.11).

Morphometry

In both PCAs (covariance and correlation matrix

approach), two separate clusters resulted from PCA with

the data subset for males that corresponded to the two

parental species (Fig. 4A). The first two principal compo-

nents based on a correlation matrix explained a smaller

percentage of the total variation (61.7%) but had slightly

higher eigenvalues (1.43, 1.05) compared to the covari-

ance matrix approach (98.2%, eigenvalue PC1 = 1.17,

PC2 = 0.17; Table S1). Factor loadings were similar in the

two analyses for males and females: PC1 was most heavily

loaded by wing span, and PC2 was most heavily loaded

by body length (except for PCA with females based on

the correlation matrix, compare Table S1). In the scatter-

plots of PC1 vs. PC2, male phenotypical house sparrows

and Spanish sparrows represented two largely separate

clusters. Most P. hispaniolensis had positive values of

PC1, whereas most P. domesticus had negative values and

male phenotypical hybrids had an almost 100% overlap

with the house sparrow cluster (Fig. 4A and C). The

dimension of PC1 reflected that on average Spanish spar-

rows had a greater wing span compared to house spar-

rows and hybrids (Table 2), and unlike for all other

biometric variables, the interaction phenotype:sex was sig-

nificant for wing span (F = 7.61, P < 0.001). The separa-

tion of the two parental species was less well reflected by

PCA results from the female data subset (Fig. 4B and D),

but differences between the parental and the hybrid phe-

notypes in scores of PC1 and PC2 were significant for

both sexes (for males: F = 166.58, P < 0.001; for females:

F = 16.47, P < 0.001). Spanish sparrows differed signifi-

cantly from house sparrows and hybrids (post hoc com-

parisons, Tukey’s test, P < 0.001), whereas the latter two

did not differ neither in males nor in females (Tukey’s

test, P = 0.71 for males and P = 0.13 for females).

Mitochondrial DNA

The haplotype network was divided into two clusters sep-

arated by a minimum of 27 substitutions (Fig. 5). Most

Table 1. Results of nest surveys from a 2-year period in two urban and one rural sparrow populations in Algeria (means for clutch size, laying

and hatching dates, incubation time, number of hatched chicks (= Young), and breeding success �SD).

Year Clutch no (n) Laying (day)

Hatching

(day)

Incubation

(days) Eggs (n) Young (n) Success (%)

P. domesticus + hybrids

Djelfa, urban

2011 1 (21) 83.7 � 2.9 96.7 � 2.9 13.0 � 0.0 4.0 � 0.8 3.4 � 0.8 85.4 � 17.2

2 (18) 119.4 � 3.5 132.8 � 3.9 13.3 � 0.5 3.5 � 0.8 3.2 � 0.9 91.0 � 17.6

3 (19) 158.1 � 8.1 171.3 � 8.1 13.3 � 1.0 3.8 � 0.9 3.1 � 0.9 82.7 � 20.5

Total n = 58 3.8 � 0.8 3.2 � 0.8 86.2 � 18

2012 1 (14) 86.0 � 4.6 99.3 � 4.8 13.3 � 4.7 3.9 � 0.7 3.4 � 0.6 86.9 � 16.5

2 (15) 120.7 � 4.5 134.1 � 4.8 13.5 � 0.6 3.9 � 0.6 3.4 � 0.5 84.8 � 19.9

3 (16) 166.1 � 5.8 179.8 � 5.4 13.7 � 0.6 3.9 � 0.8 3.3 � 0.9 84.1 � 17.9

Total n = 45 3.9 � 0.7 3.3 � 0.7 85.2 � 18

P. domesticus + hybrids

Hassi El-Euch, urban

2011 1 (18) 84.4 � 3.1 97.3 � 3.1 12.9 � 0.2 4.1 � 0.8 3.6 � 0.7 88.7 � 12.0

2 (11) 112.1 � 2.3 125.1 � 2.5 13.0 � 0.4 3.7 � 0.8 3.3 � 1.0 87.6 � 21.0

3 (14) 157.9 � 4.9 171.9 � 4.8 14.0 � 0.4 3.9 � 0.6 3.3 � 0.6 86.5 � 16.8

Total n = 43 4.0 � 0.7 3.4 � 0.8 87.7 � 16

2012 1 (13) 85.9 � 2.6 97.9 � 2.3 12.0 � 0.6 4.1 � 0.5 3.5 � 0.7 86.2 � 17.5

2 (9) 117.7 � 4.1 130.9 � 4.2 13.2 � 0.4 3.7 � 0.7 3.0 � 0.5 83.5 � 16.4

3 (9) 161.1 � 3.3 174.9 � 3.5 13.8 � 0.4 3.8 � 0.4 3.4 � 0.7 93.5 � 13.0

Total n = 31 3.9 � 0.7 3.4 � 0.7 87.6 � 16

P. hispaniolensis

Hassi El-Euch, rural

2011 1 (79) 110.0 � 2.9 123.0 � 2.9 13.0 � 0.0 4.0 � 0.8 3.2 � 0.7 80.5 � 16.0

2 (45) 154.4 � 3.9 168.2 � 4.0 13.8 � 0.4 4.0 � 0.9 3.4 � 0.7 84.5 � 14.5

Total n = 124 4.0 � 0.8 3.2 � 0.8 82.0 � 16

2012 1 (84) 114.0 � 2.4 127.8 � 2.5 13.8 � 0.4 3.9 � 0.7 3.1 � 0.7 81.2 � 16.4

2 (43) 156.2 � 3.8 169.7 � 4.3 13.5 � 0.7 3.9 � 0.8 3.2 � 0.8 83.1 � 16.8

Total n = 127 3.9 � 0.8 3.2 � 0.7 81.8 � 17



tip haplotypes differed by only one substitution from the

most common haplotype of each cluster. The house spar-

row cluster comprised 35 P. domesticus haplotypes, the

most common central haplotype dom1 (n = 77 individu-

als) was found in all house sparrow populations and was

also highly abundant in Algerian populations of Spanish

sparrows and mixed urban populations. However, haplo-

type dom1 was nearly absent from P. italiae populations

from the Italian Peninsula, Sicily and Ustica, whereas

haplotype dom2 was the most abundant one there

(Figs 5, S3). All other Italian haplotypes were unique vari-

ants derived from dom2 and did not occur in other study

populations (Fig. 5). Despite being phenotypically close

to Spanish sparrows, all birds from Lampedusa carried

P. domesticus haplotypes, too. However, unlike in other

P. italiae populations haplotype dom1 was the most

Figure 4. Morphometric variation among

North African sparrows and their phenotypical

hybrids in males (A) covariance matrix, (C)

correlation matrix and females, (B) covariance

matrix, (D) correlation matrix. Results are from

principal component analysis (PCA) of four

morphological parameters (body length, wing

span, beak length, and tarsus length)

Table 2. Means (�SD) of body size parameters for Algerian house sparrows (P. domesticus), Spanish sparrows (P. hispaniolensis) and their

hybrids.

N Weight (g)

Body length

(cm) Wing (cm) Beak (cm) Tarsus (cm)

P. hispaniolensis

Male 63 26.7 � 1.0 15.0 � 0.4 25.3 � 0.8 1.16 � 0.04 3.27 � 0.12

Female 24 25.7 � 1.0 14.8 � 0.4 24.6 � 1.0 1.17 � 0.04 3.13 � 0.12

P. domesticus

Male 71 26.3 � 0.8 14.8 � 0.4 23.4 � 0.4 1.19 � 0.06 3.21 � 0.16

Female 24 25.7 � 0.7 14.7 � 0.4 23.4 � 0.8 1.17 � 0.06 3.13 � 0.14

Hybrids

Male 25 26.2 � 0.8 14.8 � 0.4 23.5 � 0.4 1.19 � 0.05 3.20 � 0.12

Female 38 25.1 � 0.6 14.7 � 0.4 23.8 � 0.5 1.17 � 0.04 3.11 � 0.16



abundant on Lampedusa island and dom2 was the rare

variant there (Fig. 5). The Spanish sparrow cluster

included 21 P. hispaniolensis haplotypes: the most com-

mon central one his1 (n = 26 individuals) was found in

all Spanish sparrow populations, and at low abundances,

it was also present in Algerian mixed urban populations

(Figs 5, S1).

MtDNA profiles of European sparrow populations (ex-

cept Italy) were in 100% accordance with the local phe-

notype: House sparrow populations included only

P. domesticus haplotypes, whereas in Spanish sparrow

populations, only P. hispaniolensis haplotypes were found

(Fig. 5). Accordingly, mismatch distributions for these

four populations (Fig. 6) were unimodal and diversity

indices were relatively low (Table 3; with lowest values

for haplotype and nucleotide diversity in Spain (P. his-

paniolensis) and Norway (P. domesticus). The same was

true for the limited samplings from Morocco (house spar-

rows only) and from Egypt (Spanish sparrows only;

Figs 5, 6, Table 3). Unimodal mismatch distributions of

Italian peninsular and island populations of P. italiae

were extremely steep due to very low intraspecific genetic

diversity. This was confirmed by low values for haplotype

diversity (h) and nucleotide diversity (p) of Italian popu-

lations (0.18 < h < 0.38; 0.00026 < p < 0.00057) com-

pared to much higher values for all house sparrow and

Spanish sparrow populations (0.52 < h < 0.93;

0.0081 < p < 0.003; Table 3). Among all study popula-

tions, only house sparrows of Sevilla showed a signifi-

cantly negative Tajima’s D (Table 3). McDonald–Kreitman test with the parental species data set, however,

did not confirm a general deviation from neutrality for

the ND2 fragment analyzed: 22 synonymous substitutions

vs. 1 nonsynonymous substitution at 31 polymorphic

sites, neutrality index, NI (Jukes–Cantor) = 0.714,

P = 0.696.

Unlike European populations, almost all Algerian pop-

ulations had a bimodal mismatch distribution due to the

local co-occurrence of both P. domesticus and P. hispan-

iolensis haplotypes (Fig. 6). Their mtDNA profiles were

not in accordance with the existing phenotypes, except

for the phenotypical house sparrows from Hassi El-Euch

city that exclusively carried P. domesticus haplotypes

(Fig. 5). Generally, all Algerian populations were

admixed, but house sparrow mtDNA was locally most

abundant regardless of phenotypical variation (Figs 5, 6).

Figure 5. Distribution and frequency of ND2 haplotypes in European and Algerian study populations of house sparrows (dom), Spanish sparrows

(his), Italian sparrows (ita), and North African hybrids (hyb). The haplotype network based on a 707-bp fragment of the mitochondrial ND2 gene

shows P. domesticus and P. hispaniolensis as two genetic clusters, separated by a minimum of 27 substitutions. Numbers in the network indicate

the numbers of individuals sharing the three most common haplotypes. Pie charts show the frequency distribution of haplotypes for each

population (according to color code of the network); Algerian study populations: DJE = Djelfa, HEE = Hassi El-Euch.



For example, 75% of the rural, phenotypical Spanish

sparrows at Hassi El-Euch carried a P. domesticus haplo-

type. Likewise, most house sparrows and phenotypical

hybrids in urban populations carried a P. domesticus hap-

lotype, but a low percentage of individuals at Djelfa car-

ried the most common P. hispaniolensis haplotype. Due

10 20 30

0.1

0.2

0.3

10 20

0.2

0.3

0.1

30

Oslo, Norway*P. domesticus

Dresden, GermanyP. domesticus

Pairwise differences Pairwise differences

Freq

uenc

y

Freq

uenc

y Algeria, DjelfaP. domesticus

Algeria, Hassi El-EuchP. domesticus Expected

Observed

20 30 40

0.1

0.2

0.3


Freq

uenc

y

Freq

uenc

y

0.2

0.4

0.6

10 10 20 30

0.8

Sardinia, Italy*P. hispaniolensis

Sevilla, SpainP. hispaniolensis

10 20 30

0.1

0.2

0.3

10 20

0.2

0.3

0.1

30


Freq

uenc

y

Freq

uenc

y

0.4

10 20

0.2

0.4

0.6

Freq

uenc

y

155

EgyptP. hispaniolensis Expected

Observed


Freq

uenc

y

0.05

0.1

0.15

10 20 30

0.2

40

0.25Algeria, Hassi El-EuchP. hispaniolensis

Central Italy*P. italiae

10 20 25

0.2

0.4

0.6

Pairwise differences

Freq

uenc

y

155

Sicily, MalettoP. italiae


Freq

uenc

y

0.2

0.4

0.6

10 20 25155

0.8

Algeria, Djelfahybrids

Algeria, Hassi El-Euchhybrids Expected

Observed

20 30 40

0.1

0.2

0.3

Pairwise differences Pairwise differencesFr

eque

ncy

Freq

uenc

y

0.05

0.1

0.15

10 10 20 30

0.2

0.4

40

0.25

Sicily, FraginesiP. italiae

10 20 25

0.2

0.4

0.6


Freq

uenc

y

155

Expected

ObservedItaly, LampedusaP. italiae


Freq

uenc

y

0.2

0.4

0.6

10 20 25155

Figure 6. Observed (dash line) and expected (solid line) mismatch distributions for European and Algerian sparrow populations; x-axis = pairwise

number of differences, y = frequency; *haplotype data set from Elgvin et al. (2011).

Table 3. Genetic diversity for European and Algerian populations of house sparrows (P. domesticus), Spanish sparrows (P. hispaniolensis) and

their hybrids; local mixed Algerian populations included domesticus and hybrids but not P. hispaniolensis.

Taxon Population n h hd p TD

P. hispaniolensis Sevilla 22 10 0.710 � 0.106 0.00165 � 0.00042 �2.12492*

Sardinia 16 7 0.883 � 0.045 0.00281 � 0.00043 �0.63940

Hassi El-Euch 26 9 0.625 � 0.109 0.01565 � 0.00370 0.64309

Egypt 7 3 0.524 � 0.209 0.00081 � 0.00036 �1.23716

P. domesticus Dresden 17 7 0.713 � 0.109 0.00185 � 0.00046 �1.58056

Oslo 16 5 0.667 � 0.113 0.00153 � 0.00034 �0.31696

Hassi El-Euch 10 7 0.911 � 0.077 0.00239 � 0.00039 �0.84270

Djelfa 28 11 0.772 � 0.077 0.00733 � 0.00324 �1.56949

Morocco 6 5 0.933 � 0.122 0.00312 � 0.00102 �0.93169

Phenotypical hybrids Hassi El-Euch 9 6 0.833 � 0.016 0.01684 � 0.00668 0.04907

Djelfa 15 5 0.676 � 0.105 0.01079 � 0.00524 �0.84678

P. italiae C Italy 15 3 0.362 � 0.145 0.00054 � 0.00023 �1.00161

Fraginesi 10 3 0.378 � 0.181 0.00057 � 0.00029 �1.40085

Maletto 11 2 0.182 � 0.144 0.00026 � 0.00020 �1.12850

Ustica 9 2 0.222 � 0.166 0.00031 � 0.00024 �1.08823

Lampedusa 15 3 0.362 � 0.145 0.00054 � 0.00023 �1.00161

N, sample size; h, number of haplotypes; hd, haplotype diversity � SD; p, nucleotide diversity � SD; TD, Tajima’s D (*Significance level P < 0.05).



to strong mtDNA admixture, Algerian populations did

not greatly differ in haplotype diversity from European

and other North African populations, but had eight to

ten times higher nucleotide diversity (Table 3).

Discussion

While there is increasing information available on the

ecological and genetic differentiation of the Italian hybrid

form, P. italiae, from the two parental species (Her-

mansen et al. 2011, 2014; Elgvin et al. 2011; Tu 2012;

Trier et al. 2014), comparable knowledge from North

Africa is limited. In this context, we document substantial

differences between European and Algerian populations

for the first time with respect to (1) the local spatial dis-

tribution of the three forms and (2) the genetic constitu-

tion of populations. Field observations on the species’

ecological preferences and breeding biology might help

explaining these differences.

Habitat choice

In our Algerian study populations, house sparrows bred in

urban areas, whereas Spanish sparrows tended to avoid

human settlements and were strongly restricted to culti-

vated areas. These different ecological preferences of the

two sparrow species in local sympatry are in good accor-

dance with the situation in other parts of North Africa

(Meise 1936; Summers-Smith and Vernon 1972; Riss

1989) and on the Iberian Peninsula (Alonso 1986; Murgui

2011; Tu 2012). Generally, ecological segregation of the

two sparrow species might be more depending on nest site

choice than on adaptation to different food resources,

because dietary preferences of house sparrows and Spanish

sparrows largely overlap on the European continent (Sum-

mers-Smith 1988; Bernis 1989; Cramp and Perrins 1994;

Haffer and Hudde 1997; Cordero 2004) and in North

African populations (El Kharrim et al. 1997; Belkacem

et al. 2012). Differences in bill dimensions that may indi-

cate adaptation to different food niches were not found in

our Algerian study populations. However, bill size dimen-

sions in sparrows are strongly influenced by abiotic factors

(e.g., precipitation regimes) and, apparently, the degree of

parental genetic contribution does not affect the large phe-

notypic variation of bills (Eroukhmanoff et al. 2013).

In accordance with previous records from Algeria and

Tunisia (Summers-Smith and Vernon 1972), hybrid indi-

viduals were always found in urban habitats, associated

with house sparrows (P. domesticus) and like the latter

they occupied nesting holes in human buildings. It seems

that habitat choice of North African hybrids strongly

depends on the presence of either of the parental species,

because in the absence of Spanish sparrows, large hybrid

sparrow populations were found in oases agrosystems

such as date palm groves, but the preference for nest sites

on human buildings remained unaltered in these hybrid

colonies (Guezoul et al. 2010, 2011; Guezhoul et al.

2013). Shared habitat preferences between hybrids and

one of the parental species have also been reported in

another well-studied hybrid zone of great tits and Japa-

nese tits (Parus major, P. minor) in Far East (Kvist et al.

2003, Kvist and Rytk€onen 2006). There, the western spe-

cies, P. major, and the similar-sized hybrids occupy the

towns whereas the smaller Asian, P. minor, breeds outside

the human settlements (Nazarenko et al. 1999; P€ackert

et al. 2005; Fedorov et al. 2009). Strikingly, the great tit

and the house sparrow are synanthropic species that are

well adapted to human settlements. Their radiation and

range expansion were strongly associated with man-made

structures (a recent eastward expansion of great tits along

the Trans-Siberian Railway; Kapitonova et al. 2011) or

with human commensalism in the sparrows (Sætre et al.

2012). Like in the great tit example, hybridization among

the two sparrow species in North Africa is very likely a

quite recent process. Before 1900, only rare and isolate

records of hybrid individuals existed east of 2°E. How-

ever, in the first quarter of the 20th century, hybrids

records steadily increased in frequency and extension

range, likely as a result of the successive arrival of house

sparrows to the cities (Summers-Smith and Vernon

1972). As a consequence, the successful adaptation urban

habitats by one species might have involved selective

advantages in several life history traits compared to their

rural counterparts (Møller 2009). Therefore, it seems that

anthropogenic land-use change may have promoted

hybridization and hybrid range expansion in North Afri-

can sparrows like it was suggested for the origin and

rapid dispersal of the hybrid Italian sparrow (Hermansen

et al. 2011).

Breeding phenology

Compared to Algerian mixed populations of house spar-

rows and hybrids in the urban areas, Spanish sparrows bred

later and raised only two clutches instead of three. In other

Algerian populations of pure hybrids, breeding phenology

was very similar to those from our urban study popula-

tions: Courtship began in early February and the first of

three successive broods started in mid-March (Guezoul

et al. 2011). In local study populations on the Iberian

Peninsula, the two parental sparrow species started first

broods at almost identical dates, but house sparrows had

more successive clutches, and thus longer breeding periods

than Spanish sparrows (Alonso 1983). In contrast, in other

populations, Spanish sparrows bred up to 13 days later

than house sparrows (Tu 2012). The shorter breeding



period of Spanish sparrows compared to that of house

sparrows in Algerian and European populations might well

be related to the migratory or nomadic behavior of the spe-

cies on the European continent (Haffer and Hudde 1997

and references therein) and North Africa (Bachkiroff 1953;

Bortoli 1973; Cramp and Perrins 1994; Haffer and Hudde

1997; White et al. 2013). Migratory behavior of Spanish

sparrows may also explain the observed differences in body

size parameters found between the two study species (i.e.,

greater wing span in P. hispaniolensis than P. domesticus),

because long and pointed wings in relation to other mor-

phometric parameters can indicate a greater migratoriness

of populations (P�erez-Tris and Telleria 2001; Baldwin et al.

2010; Nowakowski et al. 2014; Hahn et al. 2016; but see

M€onkk€onen 1995). It has been suggested that even minimal

differences in the timing of breeding phenology may act as

premating isolation mechanism and prevent hybridization

between house sparrows and Spanish sparrows (Tu 2012).

This explanation, however, seems unsatisfying for our study

area, because the breeding periods of the two species greatly

overlap and both have more than one clutch. Furthermore,

although the breeding period of Spanish sparrows started

about 1 month later than that of house sparrows, a consid-

erable number of hybrids were present in our study popula-

tions and mtDNA introgression – although largely

unidirectional – was apparent even in the two parental spe-

cies. Thus, further explanations for such considerable

degree of mitochondrial gene flow have to be considered.

Molecular genetics – Asymmetricmitochondrial introgression in North Africa

Local compositions of mitochondrial gene pools suggest

different demographic histories of sparrow hybrid popula-

tions on the Italian Peninsula and on the North African

continent, respectively. While in P. italiae the house spar-

row mitogenome is near-fixed (Elgvin et al. 2011; Her-

mansen et al. 2011), a small percentage of Algerian hybrids

carried P. hispaniolensis haplotypes. Furthermore, we

observed strong mitochondrial introgression of house spar-

row mitogenome into the North African P. hispaniolensis

populations. The small percentage of P. hispaniolensis hap-

lotypes in the urban house sparrow population at Djelfa

shows that mitochondrial introgression is largely but not

completely unidirectional in these North African popula-

tions. Thus, we would assume that despite spatial separa-

tion hybridization between the two sparrow species in

Algeria is not a historical process, but still ongoing. The

extent and the directionality of genetic introgression across

a hybrid zone can be generally affected by complex interac-

tions of sexual selection (Helbig et al. 2001; While et al.

2015) and natural selection, such as differential adaptation

to local environments (Rheindt 2011; Walsh et al. 2016).

Furthermore, there is mixed evidence that niche divergence

itself might promote asymmetrical introgression among

hybridizing species (Wielstra and Arntzen 2014; Jim�enez

and Ornelas 2016). Hermansen et al. (2011) argued that

hybridization among house sparrows, Spanish sparrows,

and their hybrids was limited to regions of low population

densities such as the contact zone of house sparrows and

Spanish sparrows in the Alps and in North Africa and the

Cape Verde Islands, where one of the species is supposedly

rare. For such numerically imbalanced populations, the

desperation hypothesis (Hubbs 1955) predicts that

restricted mate choice in the rare species should increase

the possibility heterospecific matings and promote

hybridization processes. Although empirical support for

Hubb’s principle has been found in a number of species,

most of them dealt with single hybrid individuals (Beier

et al. 1997; McCracken and Wilson 2011; Ralston et al.

2015), and thus, this theory does not serve as a general rule

(for counterevidence see Randler 2008). In fact, field data

from North Africa showed that none of the three sparrow

phenotypes is actually rare (Guezoul et al. 2010, 2011;

Bendjoudi et al. 2013), but nevertheless, the spatial separa-

tion among populations from urban and rural environ-

ments might enhance mixed matings in a way that house

sparrows (and hybrids) are rare in rural Spanish sparrow

populations and vice versa. Unfortunately, the actual per-

centage of mixed matings in North African sparrow popu-

lations has not been estimated yet and due to the

problematic identification of female phenotypes that will

remain a challenge to future field studies.

Molecular genetics – Diversity loss andgenetic drift in the Italian sparrow, Passeritaliae

For the origin of the Italian hybrid form, P. italiae, Her-

mansen et al. (2011) developed a recent Holocene sce-

nario and assumed hybridization processes to be strongly

linked to the intensification of agriculture at about

10,000 years ago (compare Sætre et al. 2012). Undoubt-

edly, the mitochondrial gene pool of the house sparrow

(including variants dom1 and dom2) must have diversi-

fied quite recently, but a late Pleistocene scenario for the

origin of the Italian hybrid is a plausible alternative for

some reasons. First, the Italian Peninsula is one of three

classical Southern European glacial refugia (along with

the Iberian Peninsula and the Balkan Peninsula; Schmitt

2007; Stewart et al. 2010) from where different genetic

lineages of some bird species originated and expanded

their ranges after glacial retreat (Brito 2005; Brambilla

et al. 2008). In contrast to the flanking regions to the

West and the East, it seems that the Italian Peninsula

constituted a more restricted and isolated refugium



(particularly when compared to the Balkans) because Ital-

ian genomes of phylogeographically structured species

rarely populated Central and Northern Europe as the Alps

acted as strong barrier to gene flow (Hewitt 2000). Statis-

tical tests of selective neutrality did not indicate a strong

signal of selection or population growth in populations

from the classical refugia except for the Iberian popula-

tion of P. hispaniolensis. However, in particular the ances-

tral P. italiae populations might have also been

bottlenecked in their Italian refugium according to their

extremely left-skewed mismatch distributions and com-

paratively low genetic diversity. A similar loss of genetic

diversity in Italian glacial refuges was confirmed for other

terrestrial species (Deffontaine et al. 2005). Furthermore,

although Elgvin et al. (2011) already found that Italian

haplotypes do not form separate haplogroups, Italian

populations differed from all our study populations in

their genetic composition with respect to the high abun-

dance of the elsewhere rare haplotype dom2 and the near

absence of the elsewhere most abundant haplotype dom1.

Such near-fixation of rare haplotypes can also be a further

possible effect of genetic drift in Pleistocene refugia

(Bayard deVolo et al. 2013). Thus, the extant genetic

diversity observed in P. italiae is likely the result of mito-

chondrial capture and near-complete fixation of rare

house sparrow haplotypes during past hybridization pro-

cesses in glacial Italian refuges. Similar scenarios were

developed for Mediterranean gulls (Liebers et al. 2004,

2010) and the Carpathian newt (Babik et al. 2005; Zie-

li�nski et al. 2014). The closer resemblance of the P. italiae

population from Lampedusa (with dom1 being the most

abundant haplotype) with P. domesticus and Algeria

hybrid populations is likely due to the close proximity of

the island to the African continent from where founder

populations might have colonized the island. This under-

lines the complex phylogeographic pattern of Mediter-

ranean hybrid sparrow populations, because despite their

strong phenotypical similarity to Spanish sparrow popula-

tions from Sicily and Lampedusa, they greatly differ with

respect to frequencies of the two most common house

sparrow haplotypes.

Conclusion

In the North African agricultural landscape mosaic, the

patchy distribution of the two sparrow species and their

hybrids is a unique spatial pattern that is not found else-

where across the sympatric Eurasian range of house spar-

rows and Spanish sparrows and the isolated range of the

Italian hybrid form. Despite different ecological prefer-

ences and timing of the breeding period, the North Afri-

can populations are characterized by strongly imbalanced

mitochondrial introgression of urban house sparrow

haplotypes into the rural Spanish sparrow populations.

Compared to North African phenotypical hybrid popula-

tions, local Italian gene pools of the stabilized hybrid

P. italiae are characterized by low genetic diversity and

high abundances of rare haplotypes. Such differences

between mitochondrial gene pools of Italian and North

African hybrids provide first evidence that different

demographic histories have shaped the extant genetic

diversity observed on both continents. Although mito-

chondrial genetic diversity and phylogeographic patterns

have a limited explanatory power for the study of

hybridization processes, it may be informative only for

taxon pairs, like the two sparrow species, represented by

distinct phenotypes. However, only nuclear markers allow

for unmistakable identification of genetic hybrid individ-

uals and populations and a more comprehensive recon-

struction of evolutionary scenarios, historical

demography, or colonization pathways as shown in spar-

rows (Elgvin et al. 2011; Hermansen et al. 2011, 2014;

Trier et al. 2014). Whether strongly asymmetrical mito-

chondrial introgression among North African sparrow

populations is reflected by patterns of nuclear gene flow

will set a future challenge to follow-up studies.

Acknowledgments

We are grateful to the University of Djelfa who granted an

annual travel fund for A.A.B. to perform laboratory and

collection work at Senckenberg Natural History Collec-

tions Dresden. The following colleagues helped with field

work in Algeria: A. Bouabdelli, S. Abidi, K. Zerouk, and

A. Slimani. We also thank J. Figuerola for providing blood

samples of Spanish sparrows. The Italian legal permission

for this study was allowed by the ISPRA (prot. 13079 del

22/03/2013) and Regione Siciliana (D. D. S. n. 1248/2013

del 26/03/2013). J. Martens and J. Hering kindly provided

samples from Morocco and Egypt. We are deeply grateful

to two anonymous reviewers for several key suggestions

that strongly improved a previous draft of the manuscript.

Conflict of Interest

None declared.

References

Abbot, R., D. Albach, J. W. Arntzen, J. E. Baird, N. Bierne, J.

Boughman, et al. 2013. Hybridization and speciation. J.

Evol. Biol. 26:229–246.

Aliabadian, M., C. S. Roselaar, V. Nijman, R. Sluys, and M.

Vences. 2005. Identifying contact zone hotspots of passerine

birds in the Palaearctic region. Biol. Lett. 22:21–23.Alonso, J. C. 1983. Estudio comparado de los principales

parametros reproductivos de Passer hispaniolensis y



Passer domesticus en Espana centro-occidental. Ardeola

30:3–21.

Alonso, J. C. 1985. Description of intermediate phenotypes

between Passer hispaniolensis and Passer domesticus. Ardeola

32:31–38.Alonso, J. C. 1986. On the study and distribution of the

Spanish sparrow (Passer hispaniolensis (Temm.)) in Iberia.

Int. Stud. Sparrows 13:35–45.Arnold, M., ed. 2016. Divergence with genetic exchange.

Oxford University Press, New York, USA.

Babik, W., W. Branicki, J. Crnobrnja-Isailovi�c, D.

Cogalnice�anu, I. Sas, K. Olgun, et al. 2005. Phylogeography

of two European newt species — discordance between

mtDNA and morphology. Mol. Ecol. 14:2475–2491.Bachkiroff, Y. 1953. Le moineau steppique au Maroc. Service

de la Defense des V�eg�etaux, Maroc 3:1–135.Baldwin, M. W., H. Winkler, C. L. Organ, and B. Helm. 2010.

Wing pointedness associated with migratory distance in

commongarden and comparative studies of stonechats

(Saxicola torquata). J. Evol. Biol. 23:1050–1063.Bayard deVolo, S., R. T. Reynolds, S. A. Sonsthagen, S. L.

Talbot, and M. F. Antolin. 2013. Phylogeography,

postglacial gene flow, and population history of North

American northern goshawks (Accipiter gentilis). Auk,

130:342–354.

Beier, J., B. Leisler, and M. Wink. 1997. Ein

Drossel- 9 Teichrohrs€anger-Hybride Acrocephalus

arundinaceus 9 A. scirpaceus und der Nachweis seiner

Elternschaft. J. Ornithol. 138:51–60.

Belkacem, A. A., M. Sekour, and S. Doumandji. 2012.

Effectiveness of mesh netting and nest destruction in

protection of crops against attack by Spanish sparrow Passer

hispaniolensis. Afr. J. Agric. Res. 7:4575–4580.

Bendjoudi, D., H. Chenchouni, S. Doumandjy, and J. F.

Voisin. 2013. Bird species diversity of the Mitidja Plain

(Northern Algeria) with emphasis on the dynamics of

invasive and expanding species. Acrocephalus 34:13–26.

Bernis, F. 1989. Los gorriones. Con especial referencia a su

distribuci�on y eto-ecolog�ıa en las mesetas espa~nolas.

Comunicaciones I. N. I. A. Serie Recursos naturales, no. 53. Pp.

114. Ministerio de Agricultura, Pesca y Alimentaci�on, Madrid.

Bortoli, L. 1973. Sparrows in Tunisia. Pp. 249–252 in

Kendeigh, S.C., and J. Pinowski, eds. Productivity,

population dynamcis and systematics of granivorous birds.

Polish Scientifica Publishers, Warszawa.

Brambilla, M., S. Vitulano, F. Spina, N. Baccetti, G. Gargallo,

E. Fabbri, et al. 2008. A molecular phylogeny of the Sylvia

cantillans complex: cryptic species within the Mediterranean

basin. Mol. Phylogenet. Evol. 48:461–472.Brelsford, A., B. Mil�a, and D. E. Irwin. 2011. Hybrid origin of

Audubon’s warbler. Mol. Ecol. 20:2380–2389.Brito, P. H. 2005. The influence of Pleistocene glacial refugia

on tawny owl genetic diversity and phylogeography in

western Europe. Mol. Ecol. 14:3077–3094.

Clement, P. 1999. Finches & Sparrows. Pp. 500. Christopher

Helm, London.

Clement, M., D. Posada, and K. A. Crandall. 2000. TCS: a

computer program to estimate gene genealogies. Mol. Ecol.

9:1657–1659.Clements, J. F., T. S. Schulenberg, M. J. Iliff, D. Roberson, T.

A. Fredericks, B. L. Sullivan, et al. 2015. The eBird/Clements

checklist of birds of the world: Version 6.9. Available at

http://www.birds.cornell.edu/clementschecklist/download/

(accessed 14 February 2015).

Cordero, P. 2004. Pardal Com�u Passer domesticus. Pp. 502–

503. En: J. Estrada, V. Pedrocchi, L. Brotons, S. Herrando,

eds. Atlas dels Ocells Nidificants de Catalunya. Institut

Catal�a d’Ornitologia/Lynx Edicions, Barcelona.

Cramp, S., and C. M. Perrins. 1994. Handbook of the birds of

Europe the Middle East and North Africa – the birds of the

Western Palearctic volume VIII – Crows to Finches. Oxford

Univ. Press, Oxford, New York.

Deffontaine, V., R. Libois, P. Kotl�ık, R. Sommer, C. Nieberding,

E. Paradis, et al. 2005. Beyond the Mediterranean peninsulas:

evidence of central European glacial refugia for a temperate

forest mammal species, the bank vole (Clethrionomys

glareolus). Mol. Ecol. 14:1727–1739.

Dickinson, E. C., and L. Christidis. 2014. The Howard and

Moore complete checklist of the birds of the world, 4th ed.

Volume 2: Passerines. Pp. 752. Aves Press, Eastbourne, UK.

Eck, S., J. Fiebig, W. Fielder, I. Heynen, B. Nicolai, T. T€opfer,

et al. 2011. Measuring Birds–V€ogel vermessen. Pp. 118.

Deutsche Ornithologen-Gesellschaft, Halle.

El Kharrim, K., E. A. Sehhar, D. Belghyti, A. Ahami, and P.

Aguesse. 1997. �Evaluation des d�egats sur les cultures

c�er�eali�eres �a travers l0�etude du r�egime alimentaire du

moineau domestique Passer domesticus L. dans la plaine du

Gharb (Maroc). Actes Inst. Agron. Vet. (Maroc) 17:61–66.Elgvin, T. O., J. S. Hermansen, A. Fijarczyk, T. Bonnet, T.

Borge, S. Stein, et al. 2011. Hybrid speciation in sparrows II:

a role for sex chromosomes? Mol. Ecol. 20:3823–3837.

Eroukhmanoff, F., J. S. Hermansen, R. I. Bailey, S. A. Saether,

and G.-P. Saetre. 2013. Local adaptation within a hybrid

species. Heredity 111:286–292.

Fedorov, V. V., V. L. Surin, O. P. Valchuk, L. P. Kapitonova,

A. B. Kerimov, and N. A. Formozov. 2009. Maintaining

morphological specificity and genetic introgression in

populations of the great tit Parus major and the Japanese tit

P. minor in the Middle Amur region. Russ. J. Genet.

45:771–780.

Gill, F. B. 2014. Species taxonomy of birds: which null

hypothesis? Auk 131:150–161.

Guezhoul, O., H. Chenchouni, M. Sekour, L. Ababsa, K.

Souttou, and S. Doumandji. 2013. An avifaunal survey of

mesic manmade ecosystems “Oases” in Algerian hot-

hyperarid lands. Saudi J. Biol. Sci. 20:37–43.

Guezoul, O., M. Sekour, K. Souttou, and S. Doumandji. 2010.

Estimated damage caused by the hybrid sparrow Passer



http://www.birds.cornell.edu/clementschecklist/download/

domesticus x P. hispaniolensis on dates (Phoenix dactylifera)

in two palm groves in Ouargla. Leb. Sci. J. 11:3–9.

Guezoul, O., H. Chenchouni, and S. Doumandji. 2011.

Breeding biology in hybrid sparrow

P. domesticus 9 P. hispaniolensis in Northern Algerian

Sahara: case study of Biskara date palm-groves. J. Adv. Lab.

Res. Biol. 1:14–21.

Gustincich, S., G. Manfioletti, G. Del Sal, C. Schneider, and P.

Carninci. 1991. A fast method for high-quality genomic

DNA extraction from whole human blood. Biotechniques

11:298–300. 302.

Haffer, J. 1989. Parapatrische vogelarten der palaearktischen

region. J. Ornithol. 130:475–512.

Haffer, J., and H. Hudde. 1997. Passer domesticus –Haussperling, Passer hispaniolensis Weidensperling. Pp. 303

in U.N. Glutz von Blotzheim, and K.M. Bauer, eds.

Handbuch der V€ogel Mitteleuropas. Vol. 14/1. Aula-Verlag,

Wiesbaden.

Hahn, S., F. K€orner-Nievergelt, T. Emmenegger, V. Amrhein,

T. Cs€org€o, A. Gursoy, et al. 2016. Longer wings for faster

springs – wing length relates to spring phenology in a long-

distance migrant across its range. Ecol. Evol. 6:68–77.Helbig, A. J., M. Salomon, S. Bensch, and I. Seibold. 2001.

Male-biased gene flow across an avian hybrid zone: evidence

from mitochondrial and microsatellite DNA. J. Evol. Biol.

14:277–287.Hennache, A., P. Rasmussen, V. Lucchini, S. Rimondi, and E.

Randi. 2003. Hybrid origin of the imperial pheasant

Lophura imperialis (Delacour and Jabouille, 1924)

demonstrated by morphology, hybrid experiments, and

DNA analysis. Biol. J. Linn. Soc. 80:573–600.

Hermansen, J. S., S. A. Sæther, T. O. Elgvin, T. Borge, E.

Hjelle, and G.-P. Sætre. 2011. Hybrid speciation in sparrows

I: phenotypic intermediacy, genetic admixture and barriers

to gene flow. Mol. Ecol. 20:3812–3822.

Hermansen, J. S., F. Haas, C. N. Trier, R. I. Bailey, A. J.

Nederbragt, A. Marzal, et al. 2014. Hybrid speciation

through sorting of parental incompatibilities in Italian

sparrows. Mol. Ecol. 23:5831–5842.Hewitt, G. M. 2000. The genetic legacy of the Quarternary Ice

Ages. Nature 405:907–913.Hubbs, C. L. 1955. Hybridization between fishes in nature.

Syst. Zool. 4:1–20.Jim�enez, R. A., and J. F. Ornelas. 2016. Historical and current

introgression in a Mesoamerican hummingbird species

complex: a biogeographic perspective. PeerJ 4:e1556.

Johnston, R. F. 1969. Taxonomy of House Sparrows and their

allies in the Mediterranean basin. Condor 71:129–139.

Kapitonova, L. V., S. M. Smirenskii, D. S. Selivanova, V. V.

Fedorov, and N. A. Formozov. 2011. The history of ranges

of the great tit (Parus major) and Japanese Tit (Parus

minor) in the Amur Region. Biol. Bull. 38:912–926.

Kvist, L., and S. Rytk€onen. 2006. Characterization of a

secondary contact zone of the Great tit Parus major and the

Japanese tit P. minor (Aves: Passeriformes) in Far Eastern

Siberia with DNA markers. Zootaxa 1325:55–73.

Kvist, L., J. Martens, H. Higuchi, A. A. Nazarenko, O. P.

Valchuk, and M. Orell. 2003. Evolution and genetic

structure of the great tit (Parus major) complex. Proc. Royal

Soc. B. Biol. Sci. 270:1447–1454.Lavretsky, P., A. Jr Engilis, J. M. Eadie, and J. L. Peters.

2015. Genetic admixture supports an ancient hybrid origin

of the endangered Hawaiian duck. J. Evol. Biol. 28:1005–

1015.

Librado, P., and J. Rozas. 2009. DnaSP v5: a software for

comprehensive analysis of DNA polymorphism data.

Bioinformatics 25:1451–1452.

Liebers, D. A., P. de Knijff, and A. J. Helbig. 2004. The herring

gull complex is not a ring species. Proc. R. Soc. Lond. B

271:893–901.Liebers, D. A., V. Sternkopf, A. J. Helbig, and P. de Knijff. 2010.

The herring-gull complex (Larus argentatus – fuscus –cachinnans) as a model group for recent Holarctic vertebrate

radiations. Pp. 351–369. In M. Glaubrecht, ed. Evolution in

action – Case studies in adaptive radiation, speciation and the

origin of biodiversity. Springer-Verlag, Berlin, Heidelberg.

Lo Valvo, F., and G. Lo Verde. 1987. Studia della variabilit�a

fenotipica delle popolazioni italiane di passere e loro

posizione tassonomica (Passeriformes, Passeridae). Riv. Ital.

Ornithol. 57:97–110.Massetti, M. 2009. I passeri Passer spp. del complesso insulare

Siculo-Maltese, nella literature. Ecol. Urbana 21:93–98.McCracken, K. G., and R. E. Wilson. 2011. Gene Flow and

Hybridization between Numerically Imbalanced Populations

of Two Duck Species in the Falkland Islands. PLOSone 6:

e23173.

McDonald, J. H., and M. Kreitman. 1991. Adaptive protein

evolution at the Adh locus in Drosophila. Nature 351:652–654.

Meise, W. 1936. Zur Systematik und Verbreitungsgeschichte

der Haus- und Weidensperlinge, Passer domesticus (L.) und

hispaniolensis (T.). J. Ornithol. 84:631–672.Metzmacher, M. 1986. La distribution des moineaux, Passer,

en Alg�erie: observations compl�ementaires. Le Gerfaut

76:131–138.Møller, A. P. 2009. Successful city dwellers: a comparative

study of the ecological characteristics of urban birds in the

Western Palearctic. Oecologia 159:849–858.

M€onkk€onen, M. 1995. Do migrant birds have more pointed

wings? A comparative study. Evol. Ecol. 9:520–528.

Murgui, E. 2011. Gorri�on Com�un – Passer domesticus. En:

Enciclopedia Virtual de los Vertebrados Espa~noles in A.

Salvador, M. Morales, eds. Museo Nacional de Ciencias

Naturales, Madrid. http://www.vertebradosibericos.org/

Narum, S. R. 2006. Beyond Bonferroni: less conservative

analyses for conservation genetics. Conserv. Genet. 7:783–787.

Nazarenko, A. A., O. P. Valchuk, and J. Martens. 1999.

Secondary contact and overlap of Parus major and Parus



http://www.vertebradosibericos.org/

minor populations in the middle Amur river basin.

Zool. Zhournal 78:372–381 (in Russian with English

summary).

Newton, I. 2008. The migration ecology of birds. Academic

Press, London, UK.

Nowakowski, J. K., J. Szulc, and M. Remisiewicz. 2014. The

further the flight, the longer the wing: relationship between

wing length and migratory distance in Old World reed and

bush warblers (Acrocephalidae and Locustellidae). Ornis

Fenn. 91:178–186.Owczarzy, R., A. V. Tataurov, Y. Wu, J. A. Manthey, K. A.

McQuisten, H. G. Almabrazi, et al. 2008. IDT SciTools: a

suite for analysis and design of nucleic acid oligomers.

Nucleic Acids Res. 36:W163–W169.

P€ackert, M., J. Martens, S. Eck, A. A. Nazarenko, O. P.

Valchuk, B. Perti, et al. 2005. The great tit (Parus major) –a misclassified ring species. Biol. J. Linn. Soc. 86:153–174.

P�erez-Tris, J., and J. L. Telleria. 2001. Age-related variation in

wing shape of migratory and sedentary blackcaps Sylvia

atricapilla. J. Avian Biol. 32:207–213.Ralston, J., D. A. Ermacor, and J. J. Kirchman. 2015. A new

wood warbler hybrid (Oreothlypis celata 9 O. ruficapilla)

from the Adirondack mountains, New York. Wilston J.

Ornithol. 127:21–28.Randler, C. 2008. Extra-pair copulation in birds. J. Avian Biol.

37:1–5.Rheindt, F. E. 2011. Genetic introgression: an integral but

neglected component of speciation in birds. Auk 128:620–632.Riss, B. 1989. Die verbreitung des weidensperlings (Passer

hispaniolensis) in Marokko. J. Ornithol. 130:367–375.Rogers, A. J. 1995. Genetic evidence for a pleistocene

population explosion. Evolution 49:608–615.Rogers, A. J., and H. Harpending. 1992. Population growth

makes waves in the distribution of pairwise genetic

differences. Mol. Biol. Evol. 9:552–569.

Sætre, G. P., S. Riyahi, M. Aliabadian, J. S. Hermansen, S.

Hogner, U. Olsson, et al. 2012. Single origin of human

commensalism in the House Sparrow. J. Evol. Biol. 25:

788–796.Schmitt, T. 2007. Molecular biogeography of Europe:

pleistocene cycles and postglacial trends. Front. Zool. 4:11.

Stewart, J. R., A. M. Lister, I. Barnes, and L. Dal�en. 2010.

Refugia revisited: individualistic responses of species in

space and time. Proc. R. Soc. B 277:661–671.

Summers-Smith, J. D., and J. D. R. Vernon. 1972. The

distribution of Passer in Northwest Africa. The Ibis 114:259–

262.

Summers-Smith, J. D. 1988. The sparrows. Pp. 342. T. & A.D.

Poyser, Calton.

Tajima, F. 1989. Statistical method for testing the neutral

mutation hypothesis by DNA polymorphism. Genetics

123:585–595.

Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and

S. Kumar. 2011. MEGA5: molecular evolutionary genetics

analysis using maximum likelihood, evolutionary distance,

and maximum parsimony methods. Mol. Biol. Evol.

28:2731–2739.T€opfer, T. 2006. The taxonomic status of the Italian sparrow –

Passer italiae (Vieillot 1817): speciation by stabilized

hybridization? Zootaxa 1325:117–145.Trier, C. N., J. S. Hermansen, G. P. Sætre, and R. I. Bailey.

2014. Evidence for mito-nuclear and sex-linked reproductive

barriers between the hybrid Italian sparrow and its parent

species. PLoS Genet. 10:e1004075.

Tu, M. H. 2012. Fine and broad scale premating isolation

between two closely related species, the house sparrow

(Passer domesticus) and the Spanish sparrow (P.

hispaniolensis). Master of Science Thesis at the Centre for

Ecological and Evolutionary Studies. Pp. 59. University of

Olso, Norway. Available at https://www.duo.uio.no/

bitstream/handle/10852/35416/Tu_Master.pdf?sequence=1

(accessed 09.09.2013).

Walsh, J., G. W. Shriver, B. J. Losen, and A. I. Kovach. 2016.

Differential introgression and the maintenance of species

boundaries in an advanced generation avian hybrid zone.

BMC Evol. Biol. 16:65.

While, G. M., S. Michaelides, R. J. P. Heathcote, H. E. A.

MacGregor, N. Zajac, J. Beninde, et al. 2015. Sexual

selection drives asymmetric introgression in wall lizards.

Ecol. Lett. 18:1366–1375.White, M., K. Stepniewski, and M. Megalli. 2013. Passerine

migrants in Bahariya oasis, Western desert, Egypt: surveys

and habitat associations. Sandgrouse 35:2–13.

Wielstra, B., and J. W. Arntzen. 2014. Exploring the effect of

asymmetric mitochondrial DNA introgression on estimating

niche divergence in morphologically cryptic species. PLoS

One 9:e95504.

Zieli�nski, P., K. Dudek, M. T. Stuglik, M. Liana, and W.

Babik. 2014. Single nucleotide polymorphisms reveal genetic

structuring of the Carpathian newt and provide evidence of

interspecific gene flow in the nuclear genome. PLoS One 9:

e97431.

Supporting Information

Additional Supporting Information may be found online

in the supporting information tab for this article:

Figure S1. Phenotypical diagnosis of target sparrow spe-

cies in Algerian populations; A) five major plumage traits

that are distinctive for house sparrows (B: P. domesticus)

and Spanish sparrows (C: P. hispaniolensis) but are inter-

mediate to a variable degree in a considerable number of

putative hybrid individuals (D) per local population

(P. domesticus 9 P. hispaniolensis).

Figure S2. Phenotypical comparison of Mediterranean

island populations (Sicily, Ustica, Lampedusa) of the Ital-

ian hybrid form, P. italiae.



https://www.duo.uio.no/bitstream/handle/10852/35416/Tu_Master.pdf?sequence=1

https://www.duo.uio.no/bitstream/handle/10852/35416/Tu_Master.pdf?sequence=1

Figure S3. Haplotype network of European and North

African sparrow populations (P. domesticus, P. hispan-

iolensis, P. italiae and North African hybrids) based on

707 bp of the mitochondrial ND2; populations of origin

are color-coded for each haplotype.

Table S1. Results of principal component analysis (PCA)

of four biometric measurements (length of body, wing,

bill and tarsus) from male and female house sparrows

(P. domesticus), Spanish sparrows (P. hispaniolensis) and

their hybrids; eigenvalues and factor loadings for the first

two principal components (PC1, PC2) based on a covari-

ance matrix and based on a correlation matrix; % = per-

centage of the total variation explained by one

component.



north african hybrid sparrows (passer domesticus p...

Documents